Nvidia RTX 5070 could launch at CES 2025, and according to rumors – including 12 GB of VRAM – everyone already hates this GPU
Nvidia’s RTX 5070 will apparently be unveiled at CES 2025, alongside the rumored RTX 5080 and 5090 next-gen graphics cards.
This is a notable change in the way the rumors are blowing, as so far we’ve heard that only the RTX 5080 and 5090 will be the first Blackwell GPUs to be launched by Nvidia.
Wccftech has made this claim and also provides some spec reveals for the RTX 5070, after speaking to sources (who provided a leak yesterday about the RTX 5080 getting a faster VRAM). In addition, two speakers on
Wccftech believes the RTX 5070 will use the GB205 chip, but doesn’t give us a core count. We’re also told that the graphics card will work with a 192-bit memory bus and 12 GB of VRAM (GDDR7). This will be the slower version of GDDR7, 28 Gbps, which will also be on board the RTX 5090 (but the 5080 will get the 32 Gbps memory, if Wccftech’s sources are correct).
So for the RTX 5070 we’re looking at a total memory bandwidth of 672 GB/s, a lot less than the supposed 1024 GB/s of the RTX 5080, but quite a bit more than the RTX 4070 (a third more, in fact).
However, we’re told that power consumption will also increase for the RTX 5070, and it will draw 250W (compared to 200W for the RTX 4070).
Leaker Harukaze5719 fully supports these specs, while adding an alleged core count of 6,400. Another regular leaker on X, Kopite7kimi, makes a backup also the 250W TDP theory (and the 12GB VRAM claim for that matter).
GB202: 24,576 RTX 5090: 21,760 / GB202-300-A1 / 512-bit / 32GB GDDR7, 28Gbps, ~1.8TB/s / 600W TBP / 14-layer FE PCBGB203: 10,752 RTX 5080: 10,752 / GB203-4 00- A1 / 256-bit / 16GB GDDR7, 32Gbps, 1TB/s, 400W TBPGB205: 6,400 RTX 5070: TBD / GB205 based / 192-bit / 12GB GDDR7,… https://t.co/KTO6QLdKCkOctober 9, 2024
As for the time frame of arrival, remember that this refers to the first unveiling of these three Blackwell graphics cards together at CES 2025 (in theory). When it comes to sales dates, the RTX 5090 and 5080 will likely be close together, but we imagine the RTX 5070 will likely be further away given its very sudden appearance in the Blackwell launch rumors.
Analysis: Nvidia’s RTX 5070 – a sheep in wolf’s clothing GPU?
Okay, deep breath. As you can imagine, there’s some dissatisfaction with these RTX 5070 specs, and especially the memory (and the move to the GB205 chip – the RTX 5080 will use GB203).
You can guess the sort of thing: “Nvidia is scamming PC gamers again!” “The RTX 5070 should have 16GB of VRAM, the RTX 5060 should have 12GB – this is an RTX 5060 sold as a 5070!” And so on…
Well, these people certainly have a point, but we can’t get ahead of ourselves here. First of all, we’re still in rumor land, and we don’t know what prices are associated with these GPUs either. Whatever the specs, Nvidia will hopefully make sure these graphics cards make sense in terms of performance per dollar and overall value.
Although we suspect the problem is that most people will find it difficult to believe that Nvidia will offer a reasonable value proposition, and that includes us. This would represent a fundamental change in Team Green’s attitude, and there should be a reason for it. The only option is the performance and value balance of RDNA 4, but remember that AMD will be in the mid-range. This means it might have some impact on the RTX 5070, but for the more expensive Nvidia Blackwell cards there won’t be any rivals, not remotely.
In short, we completely understand the concerns here, but let’s not judge these graphics cards based on rumors before they even arrive, as tempting as that may be.
But if it’s true that the RTX 5070 will have 12 GB, and the RTX 5080 will have 16 GB, you have to ask yourself: why? This really doesn’t feel like enough VRAM for PC gamers right now, let alone remotely future-proof gear as the years go by (and we’ll still have quite a while until RTX 6000 of course).
The cynical will obviously suggest that these video RAM configurations give Nvidia room to maneuver further with ‘Super’ or ‘Ti’ versions. Rumor has it that a 24GB variant of the RTX 5080 is coming later, which may be the RTX 5080 Super, and this will allow Nvidia to charge more for what most gamers would see as the right amount of VRAM for this level of graphics card in the current gaming climate. Perhaps this also applies to a 5070 with 16 GB of video memory, which all feels like cynical tactics to maximize revenue and profit.
This raises another specter: when we get the RTX 5060, it will probably have 8 GB? As if there weren’t enough complaints about the current RTX 4060’s retention of 8GB. Yes, we’re getting sucked into a bit of a negative spiral of worry here, but we have to keep reminding ourselves: these are just rumors. And Nvidia needs to ensure that these graphics cards make sense somehow, in terms of their overall performance and price tag.
However, it feels like Nvidia is pushing gamers who play at 4K to buy their most expensive GPUs, and you can bet those top-end Blackwell offerings will be seriously expensive.
On a final, more positive note, Wccftech says Nvidia is unveiling “next-generation AI technologies” for the PC gamer, which will be a big surprise at CES 2025. Color us intrigued. A way for DLSS 4 to make up for the meager VRAM equipment, perhaps?
Via VideoCardz